Interdisciplinary Project Led by Prof. Agnieszka Wyłomańska Receives Funding from Weave-UNISONO

An interdisciplinary project led by Dr. hab. Eng. Agnieszka Wyłomańska, full professor at the Faculty of Mathematics, has been selected as one of three proposals to receive funding in the latest round of the Weave-UNISONO competition. The National Science Centre has awarded our researcher PLN 1.3 million to advance signal processing techniques for cyclostationary models.

The project, titled “Advanced Signal Processing Techniques for Cyclostationary Models in the Presence of Gaussian and Non-Gaussian Noise – Source Detection, Estimation, Algorithm Optimization, and Validation in the Context of Fault Identification”, brings together mathematicians, experts in condition monitoring and fault detection from the mining and energy sectors, and partners from Belgium, creating a comprehensive approach to data analysis that has the potential to revolutionize technical diagnostics of machinery.

Prof. Agnieszka Wyłomańska will conduct the research together with the team of Prof. Radosław Zimroz from the Faculty of Geoengineering, Mining and Geology, and Prof. Konstantinos Gryllias from KU Leuven.

The Weave-UNISONO competition announces results on a rolling basis. Its creation is the result of a multilateral collaboration among research funding institutions united under Science Europe. The competition aims to simplify the submission and selection process for research projects across all scientific disciplines involving researchers from two or three European countries.

In this funding round, proposals were evaluated based on collaboration with research institutions in Belgium – Flanders. Only three projects from all of Poland were selected for funding, including one from Wrocław University of Science and Technology.

The Mathematical Rhythm of the World Around Us

Contemporary science allows the recording of data describing cyclical processes, such as ocean waves, breathing, or the human heartbeat. These phenomena, captured as signals, are best described by mathematical models known as cyclostationary models. Although research on these models has been ongoing for years, analyzing signals corrupted by impulsive (non-Gaussian) noise remains a challenge.

"We aim to systematize our understanding of the sources of cyclostationary processes and propose new physical and mathematical models," explains Prof. Agnieszka Wyłomańska from Faculty W13. "Within our interdisciplinary team, we will develop advanced algorithms for preprocessing, decomposing, and classifying signal segments."

Polish and Belgian researchers will employ robust dependency measures and innovative signal representations in the form of two-frequency maps.

"Although the developed algorithms will be validated using data from machinery in the raw materials and energy sectors, they will be universally applicable," says Prof. Radosław Zimroz from the Faculty of Geoengineering, Mining and Geology, who will participate in the project. "Growing demands for system reliability make precise diagnostics of rotating machinery increasingly crucial."

The research will focus on detecting local faults in rolling bearings and gearboxes using vibration signals. The new methods will allow diagnostic information to be separated from noise, significantly improving fault detection efficiency. These solutions will be tested not only in laboratory settings but also under challenging industrial conditions.

"Imagine trying to hear the ticking of a clock in the middle of a noisy renovation site," summarizes Prof. Agnieszka Wyłomańska. "The methods we are developing are like intelligent headphones that can mute the sound of a pneumatic hammer to precisely capture the regular rhythm of the mechanism, alerting us the moment it begins to operate abnormally."